Each year, millions of Americans are adversely affected by stress and trauma, which can have a serious impact on health, psychological well-being, and psychosocial adjustment (Kessler, Sonnega, Bromet, Hughes, & Nelson, 1995). The negative effects of stress and trauma include coronary artery disease, hypertension, and peptic ulcer disease (Rosenman & Chesney, 1982; Selye, 1982, 1976). in our society, certain groups are at higher risk for many of the effects of stress and trauma, including women, ethnic minorities, and the urban poor (cf. Dohrenwend, 2000). We propose to address these issues by studying a sample of urban, predominantly female, ethnically diverse students attending a public college in a large city. The goal of the proposed research is to analyze how the negative effects of stress on psychosocial adjustment can be mitigated by a particular intervention, "emotional disclosure" (cf. Pennebaker, 1997a). The intervention involves expressing one's thoughts and feelings about specific stressful events in writing during a succession of three twenty-minute sessions. This has been shown in past research to be an intervention that can lead to benefits in a range of outcome variables, but it does not benefit everyone equally. Using a two-cohort, longitudinal design, we will study the role played by both moderator and mediator variables. We will test three hypotheses about moderator variables: that men benefit from this intervention more than women, that (when participants write in English) native English speakers benefit more than non-native speakers, and that participants suffering mild to moderate distress will benefit more than those suffering either severe distress or little to no distress. Furthermore, we will test the theory that stress affects psychosocial adjustment indirectly, with self-efficacy and perceived social support serving as mediator variables, and that it is by affecting these mediator variables that emotional disclosure leads to benefits (cf. Solberg & Villareal, 1997). The emotional disclosure intervention also involves participants producing narratives about important stressors. We will use such methods as computer text analysis to study these narratives as sources of detailed information about the stresses faced by college students, and how these stresses vary depending on demographic variables (cf. Pennebaker & Francis, 1996). Given unique features of our sample, the proposed research will prove especially useful in advancing our understanding of the negative effects of stress on minorities and the urban poor, and how these effects can be mitigated with a brief, low-cost intervention.